What is a prompt for AI and how to write one?

What is a prompt for AI?

A prompt is a command or set of instructions given to an artificial intelligence model (e.g., ChatGPT or Gemini), which defines what the user wants to obtain in response. A prompt can be as simple as a question or as complex as a detailed structure describing a situation or problem. In interactions with AI, the prompt is a crucial element because the quality and relevance of the AI-generated response depend on its content. A well-crafted prompt will yield a solid answer from the AI, while a poorly structured one may lead to miscommunication.

Prompts can be categorized into several main types, depending on what the user aims to achieve. Each type has its specific application and should be tailored to the user’s goal. These can include:

Commands – e.g., "Write an article about…", "Create an outline for an article on X in 5 points," or "List the benefits of…" or "Write a code snippet for…".

Questions – e.g., "What are the main features of…" or "What is…?".

Descriptions – e.g., "Describe a scene where…" or "Describe the product…".

A different prompt leads to various AI responses. For example, consider “Tell me about artificial intelligence” versus the more specific “Explain in simple terms what artificial intelligence is and its applications in education.” In the first case, ChatGPT provided a response divided into these points:

Main aspects of artificial intelligence (machine learning, deep learning, natural language processing, conversational AI, image and video recognition).

Examples of AI applications (medicine, marketing and advertising, finance, entertainment and media, industry and manufacturing).

The future of AI.

AI-related challenges.

To learn more about each of these topics, we could continue the conversation with ChatGPT. In the second case, however, ChatGPT offered a response divided into the following points:

Applications of artificial intelligence in education (personalized learning, voice assistants and chatbots, automatic grading, educational content creation, progress analysis).

Two prompts on the nature of artificial intelligence, two entirely different types of responses — as you can see, how we ask a question to AI affects the quality and specificity of the answer we receive.

➡️ Prompts are also crucial when creating graphics — such as with DALL·E. Suppose we want to create an image for a cow's milk advertisement. Depending on the level of detail in our prompt, we’ll get images that are more or less aligned with our vision. The following two images were generated using the AI model mentioned. The prompt for the first one was “Graphic for a cow’s milk advertisement. Scene in a meadow.” The second was “Generate an image for a cow’s milk advertisement, showing the consumption of milk from a cow grazing in a green meadow.”

Of course, both images would still require a specialist’s touch — for example, with text, adjusting colors to match brand hues, reducing any overly repetitive elements, or adding the producer's logo and advertising slogan.

How to write a good prompt? Prompt-creation guidelines

As we can see, for AI to provide a satisfactory result, the prompt shall be crafted perfectly. Writing effective prompts requires following a few key principles. The most important ones include:

Precision – prompts should be specific and clear so that AI understands exactly what the user expects. Avoid vague statements or generalizations.

Context – it’s advisable to provide additional context if needed for a more relevant response. For instance, prompts about history should indicate specific time frames.

Length constraints – overly lengthy prompts can make it difficult for AI to understand the main intent, so it’s best to keep them concise while maintaining clarity.

Too much information can cause AI to… simply get confused. Remember, we’re not talking to a person, but to a machine that treats our words as variables. For example: DALL·E generated the following image based on the prompt “Generate an image without an elephant.” Drawing attention to the presence of an elephant and requesting it to be excluded — while using the word "elephant" — will likely lead to an image featuring an elephant regardless.


Ambiguity leads to misleading or off-topic responses. For example, the question “What is marketing?” is too general and results in a wide range of answers, while a more precise prompt like “What are the key features of digital marketing?” yields a more focused response. It’s important to craft prompts with a clear purpose and avoid ambiguous phrases.

➡️ To enhance prompt effectiveness, it’s helpful to:

Use clear and understandable words – avoid jargon unless necessary, and explain any terms that might be difficult to understand.

Apply instruction sequences – if the prompt includes multiple elements, consider breaking it into parts or presenting it as a numbered list. You can also request responses in parts, such as “Describe point 1” and then “...point 2.”

Test different prompt variations – rephrasing or adding new details to the prompt helps in achieving more accurate responses.

Well- and poorly-formulated prompts

A good way to understand how to create effective prompts is by analyzing specific examples. A well-formulated prompt considers not only the topic but also the scope, tone, and expected sources of information. Let’s look at two similar prompts that lead to different results:

Poorly formulated prompt: “Write an article about health.”

Well-formulated prompt: “Write an article on the impact of a plant-based diet on heart health, including studies from the past five years.”

➡️ When writing prompts, it’s helpful to follow a three-step guide:

1. Define the purpose of the prompt – think about what you want to achieve and clearly express it.
2. Add details or context – include additional information to narrow down the topic or provide key data.
3. Ask a question or give a command – depending on the goal of the prompt, use clear questions or specific commands.

If the goal is to get a list of tips on running a marketing campaign, it’s better to use a prompt like: “Provide 5 key tips for running an effective marketing campaign for B2B startups,” instead of the broad “How to run a marketing campaign?”

Tools for writing prompts and editing results

There are many tools and applications available to assist with writing prompts and editing the text that AI provides in response — remember, the AI's answer is just a starting point, and it's up to us to refine and adapt the content for our business needs. Grammar and style-checking applications are especially useful for creating more precise prompts:

Grammarly

Hemingway Editor

ProWritingAid

LanguageTool

QuillBot

Testing prompts is a crucial step in the creation process. After writing a prompt, it’s essential to test it in practice to ensure that the responses align with your goals. We recommend making small adjustments and observing how the AI responds to modifications. This process allows for fine-tuning the prompts until the responses are satisfactory.

Summary

Creating effective prompts requires precision, clarity, and practice. A well-formulated prompt not only yields more accurate and relevant responses from artificial intelligence but also greatly enhances the efficiency of working with AI tools. Remember to test different approaches and refine techniques based on results. By applying the principles described and using the available tools, we can master the skill of crafting prompts that fully leverage the potential of artificial intelligence!

➡️ Read also: Is GEO – Generative Engine Optimization – the future of SEO?